Targeting of malate synthase 1 to the peroxisomes of Saccharomyces cerevisiae cells depends on growth on oleic acid medium Markus Kunze 1 , Friedrich Kragler 1, *, Maximilian Binder 2 , Andreas Hartig 1 and Aner Gurvitz 1 1 Institut fu È r Biochemie und Molekulare Zellbiologie der Universita È t Wien and Ludwig Boltzmann-Forschungsstelle fu È r Biochemie, Vienna Biocenter, Austria; 2 Institut fu È r Tumorbiologie-Krebsforschung der Universita È t Wien, Vienna, Austria The eukaryotic glyoxylate cycle has been previously hypothesized to occur i n the peroxisomal compartment, whichintheyeastSaccharomyces cerevisiae additionally representsthesolesiteforfattyacidb-oxidation. The sub- cellular location of the key glyoxylate-c ycle enzyme malate synthase 1 (Mls1p), an SKL-terminated protein, was examined in yeast cells grown on d ierent carbon sources. Immunoelectron microscopy in combination w ith cell f rac- tionation showed that Mls1p was abu ndant in the peroxi- somes of cells grown on oleic acid, whereas in ethanol-grown cells Mls1p was primarily cytosolic. This was reinforc ed using a green ¯uorescent protein (GFP)±Mls1p r eporter, which e ntered p eroxisomes solely in cells grown und er oleic acid-medium conditions. Although growth of cells devoid of Mls1p on ethanol or acetate could be fully restored using a cytosolic Mls1p devoid of SKL, this construct could only partially alleviate t he requirement for native Mls1p in cells grown on oleic acid. The combined results indicated that Mls1p remained in the cytosol o f cells grown on e thanol, and that targeting of Mls1p to the peroxisomes was advanta- geous to cells grown on o leic acid as a sole carbon source. Keywords: Saccharomyces cerevisiae; glyoxylate cycle; peroxisome; m alate s ynthase 1; oleic acid. Microorganisms are able t o g row on nonfermentable carbon sources such as acetate, ethanol, o r fatty acids, because they possess a glyoxylate cycle for generating four- carbon units that are suitable for biosyntheses of macro- molecules. Similarly, plant seedlings can also u se stored lipids as a sole carbon and energy source, by converting the acetyl-CoA product of fatty acid b-oxidation to four-carbon units using a cognate process. In those eukaryotes known to possess a glyoxylate cycle, e.g. plant seedlings and fungi, the process is thought to occur in the peroxisomal matrix. Peroxisomes typically cont ain enzymes f or reactions involving m olecular oxygen a nd for metabolizin g hydrogen peroxide [1]. This subcellular compartment represents the site of fatty a cid b-oxidation, which in mammals is augmented by an additional p rocess found in the m ito- chondria [2]. The signi®cance of the fungal glyoxylate cycle to human health is underscored by the requirement of isocitrate lyase for the virulence of the pathogenic yeast Candida albicans [3]. Like the situation with C. albicans, Saccharomyces cerevisiae cells isolated from phagolyso- somes obtained f rom infected mammalian c ells similarly up-regulate isocitrate lyase as well a s m alate synthase, both of which represent key enzymes unique to the glyoxylate cycle [3]. As S. cerevisiae is a genetically more tractable yeast than C. albicans, it was chosen as a model fungal system for studying the glyoxylate cycle by analysing the subcellular distribution of malate synthase 1. The S. cerevisiae glyoxylate cycle (Scheme 1) consists of ®ve enzymatic activities, some of which are represented by isoenzymes: i socitrate lyase, Icl1p [4]; malate synthase, Mls1p and Dal7p [5]; malate dehydrogenase, Mdh1p [6], Mdh2p [7] and Mdh3p [8,9]; citrate synthase, Cit1p [10], Cit2p [ 11,12] and Cit3p/YPR001w [13]; and aconitase, Aco1p [14] and Aco2p/YJL200c [13]. As mentioned above, isocitrate lyase and malate synthase represent key enzyme activities that are unique to the glyoxylate cycle, whereas some of the remaining enzymes, e.g. mitochondrial Cit1p, Mdh1p, and Aco1p, are shared with the citric acid cycle. Icl1p is an extraperoxisomal protein, w hile Mdh3p and Cit2p are peroxisomal ones. The latter two enzymes end with a C-terminal SKL tripeptide representing a p eroxiso- mal targeting signal PTS1 [15±17]. The two malate synthases Mls1p and D al7p are a lso SKL- terminating p roteins that are 81% identical to one another. However, as the MLS1 gene is highly tr anscribed on nonfermentable carbon sources and is essential for cell growth on these m edia, whereas DAL7 is not [5], it is reasoned that only Mls1p represents the malate synthase activity speci®cally involved in the glyoxylate cycle. Dal7p, whose peroxisomal location re mains putative, is actually thought to be involved in the metabolism of glyoxylate produced during the degradation o f a llantoic acid t o urea [ 5]. Initial work on peroxisomal citrate synthase (Cit2p) led to the conclusion that the glyoxylate cycle is a peroxisomal process [12]. Howeve r, the c ycle's subcellular l ocation is n o longer clear because peroxisomal Cit2p has since been shown to be dispensable for the glyoxylate cycle [9] and, moreover, cells lacking peroxisomal malate dehydrogenase Correspondence to A. Hartig, Institut fu È r Biochemie und Molekulare Zellbiologie, Vienna Biocenter, Dr Bohrgasse 9, A-1030 Vienna, Austria. Fax: + 43 1 4277 9528, Tel.: + 43 1 4277 52817, E-mail: AH@abc.univie.ac.at Abbreviations: PTS1, peroxisomal targeting signal type 1; YP, yeast extract/peptone; GFP, green ¯uorescent protein; Mls1p, malate synthase 1; Cit2p, peroxisomal citrate synthase. *Present address: Se ction of Plant Biology, Di vision of Biological Sciences, U niversity of California, One S hields Avenue, Davis, CA 95616, USA. (Received 2 August 2001, revised 3 December 2001, accepted 5 December 2001) Eur. J. Biochem. 269, 915±922 (2002) Ó FEBS 2002 (Mdh3p) grow abundantly on ethanol [18]. I nstead, t he malate dehydrogenase activity speci®cally involved in the glyoxylate cycle is attributed to the cytosolic isoform Mdh2p [7]. The suggestion of an extra-peroxisomal location for the yeast g lyoxylate cycle was further reinforced by the demonstration t hat Icl1p is a cytosolic enzyme [4], an d that pex mutants lacking functional p eroxisomes grow plentifully on ethanol as sole carbon source [19]. The present work was aimed at determining the subcellular location of the glyoxylate cycle by examining the partitioning of Mls1p in cells grown on media supplemented with ethanol or oleic acid. MATERIALS AND METHODS Strains, plasmid constructions and gene disruptions S. ce revisiae strains, plasmids and o ligonucleotides used are listedinTable1.Escherichia coli strain HB101 was used for all plasmid ampli®cations and isolations. Construction of strains JD1, JR85, and JR86 has been described [5]. To remove the three codons for SKL from the MLS1 gene, single-strand mutagenesis was performed according to the manufacturer's protocol (Amersham Pharmacia B iotech., Stockholm, Sweden) using oligonucleotide H161 ( Table 1). To reintroduce the native MLS1 or an MLS1 variant lacking the SKL codons back to the genomic MLS1 locus, strain JR86 was transformed with URA3-marked integra- tive plasmids pB10-WT or pB10-WT DSKL digested with PvuII. These pUC18-based plasmids consisted of the promoter and terminator regions of MLS1 delineating the open read ing frame, with or without the codons for SKL, and URA3 (Scheme 2). Integration of the disruption fragments resulted in the respective strains KM10 and KM11. Correct integration of t hese plasmid fragments was veri®ed by p olymerase c hain reaction using oligonucleotide pairs H338 and H162, or H339 and H 161, respectively (Table 1, Scheme 2). To generate n ull mutants devoid of Mls1p, the corre- sponding gene was deleted by transforming strains BJ1991 [20] with an mls1D::LEU2 disruption fragment [5]. Cells that had r eturned to l eucine prototrophy were veri®ed for growth de®ciency on ethanol and acetate media and were designated strain KM12. The mutant phenotype was con®rmed by complementation using native MLS1 carried on a YEp352 multicopy vector, YEp352-MLS1 [5]. The BJ1991-derived strain KM13 expressing the SKL-less Mls1p was constructed a nd veri®ed as described above for strain KM11. YEp352-MLS1DSKL was constructed by inserting a 2.3-kb SalI fragment containing the complete MLS1 gene into this multicopy vector, and replacing parts of the c oding region with the single-strand mutagenized sequence, resulting in the expression of an SKL-truncated Mls1p (Mls1pDSKL). The plasmid was introduced to strain JR86, resulting in strain KM15. To create a reporter construct based on GFP extended by the C-terminal half of Mls1p comprising 274 amino acids of a total of 554, PCR was applied to YEp352-MLS1 template DNA using oligonucleotides H623 and H625 and Pfu high- ®delity polymerase (Stratagene, La Jolla, CA, USA). The single ampli®cation pro duct obtained w as digested with SphIandBglII, and ligated to an SphI- and BamHI-digested plasmid pJR233M [21], resulting in plasmid pLW89. Construction of the parent plasmid pJR233 is described elsewhere [22]. Nucleic acid manipulations [23] and y east transformations [24] were performed as described. Media and growth conditions Plates contained 0.67% (w/v) yeast nitrogen base without amino acids (Difco), 3% (w/v) agar, amino acids as required, and either 2% (w/v) D -glucose, 2.5% (v/v) ethanol, or 0.1 M potassium acetate at p H 6.0. Fatty acid plates contained 0.125% (w/v) oleic acid, and 0.5% (w/v) Tween 80 to emulsify the fatty acids [25], but lacked yeast extract. For oleic acid utilization assays and cell fractiona- tions, cells were grown overnight in rich-glucose medium consisting of YP (1% w/v yeast extract, 2% w/v peptone) and 2% D -glucose, transferred to YP containing 0.5% D -glucose at a 1 : 1 00 dilution, and grown to late log phase. Cells were transferred t o water at a concentration of 10 4 cellsámL )1 , serially diluted (1 : 10 dilutions), and culture aliquots of 2.5 lL were applied to solid media [25,26]. Growth assays in liquid oleic acid medium were performed following a modi®ed protocol [25,26]. Cells were grown overnight in synthetic medium (0.67% yeast nitrogen b ase with amino acids added) containing 2% D -glucose, and the cultures diluted to a n D 600 of 0.5 in synthetic medium containing 0.5% D -glucose and grown further with s haking at 30 °C. Upon reaching an D 600 of 3.0 culture aliquots were removedanddilutedtoanD 600 of 0.02 in synthetic media containing 0.03 M potassium phosphate buffer (pH 6.0), 0.1% yeast extract, and either 2% ethanol or 0.2% oleic acid and 0.02% Tween 80 (the latter carbon s ource adjusted prior Scheme 1. The glyoxylate cycle in yeast cells grown on ethanol. To synthesize sugars from C 2 carbon sources, yeast c ells rely on the gly- oxylate cycle. This process is based on some of the same enzymes as those of the citric acid cycle. H owever, the steps in whic h decarboxy- lations occur in the latter cycle are bypassed using two glyoxylate-cycle speci®c enzymes, isocitrate lyase a nd mal ate sy nthase. The S. cerevisiae enzymes Icl1p, Mls1p, Mdh2p, Cit1p, and Aco1p are noted, these being essential for growth o f yeast cells on C 2 carbon sources such as ethanol or acetate. 916 M. Kunze et al.(Eur. J. Biochem. 269) Ó FEBS 2002 to dilution to pH 7.0 with NaOH). The D 600 of the c ultures was determined at the times i ndicated. For vital counts, culture aliquots were removed following the i ndicated periods and plated on solid YP medium containing 2% D -glucose for enumeration following 2 days incubation. Cell fractionation and immunoblotting Late log-phase cells were harvested by centrifugation and transferred t o Y P medium containing 2.5% ethanol, or 0.2% oleic a cid and 0.02% Tween 80 (pH adjusted a s mentioned above). Following growth for a t least 9 h at 30 °C with shaking, cells were harvested by centrifugation (5000 g), and total homogenates, organellar pellets, a nd postorganellar supernatants were prepared as described [27]. A 1 0% portion of each of the f ractions (postnuclear supernatant, organellar pellet or cytosolic supernatant) was used for protein precipitation. These organellar or super- natant fractions were made u p to 0.5 mL w ith breaking buffer [27], followed by 5 lL Triton X-100 (®nal concen- tration 1% v/v) and an appropriate amount of 80% (w/v) trichloroacetic acid to obtain a 10% ®nal concentration of trichloroacetic acid. The resulting oily pellet was washed once with a diethyl ether/ethanol mixture (1 : 1), which removed traces of Triton X-100 and t richloroacetic acid, and dissolve d in 30 lL0.1 M NaOH. To the solubilized protein a volume of 30 lL sample buffer (100 m M Tris/HCl at pH 6.7; 20% w/v glycerol; 2.0% w/v SDS; 6 M urea; 100 m M dithiothreitol; and 0.1% w/v bromophenol blue) was added, and the mixture was heated to 80 °Cpriorto resolution by electrophoresis on an SDS/polyacrylamide gel (10% w/v) [28]. Following electrophoresis, the resolved proteins were transferred to a nitrocellulose ®lter according to a standard protocol. D etection o f the immobilized proteins was performed by adding a primary antibody against Mls1p (diluted 1 : 2000) or peroxisomal catalase A (Cta1p, diluted 1 : 1000) [27], followed by application of the enhanced chemiluminescence (ECL) system from Pierce (Super Signal West Pico Chemiluminiscent Substrate; no. 34083). Determination of protein concentration w as per- formed as described [29]. Puri®cation of tagged Mls1p and generation of anti-Mls1p Ig To obtain pure protein for generating an antibody against Mls1p, the pQE-32 expressio n s ystem (Qiagen Inc., V alencia, CA, USA) was used. A DNA fragment encoding the Table 1. S. cerevisiae strains, plasmids, and oligonucleotides used. The n umbers in superscript follow ing t he strains' designation refer to t heir parental genotypes, e.g. JD 1 was derived from (1) GA1-8C. Strain, plasmid, or oligonucleotide Description Source or Reference Strains (1) GA1-8C MATa ura3-52 leu2 his3 trp1-1 ctt1-1 gal2 [5] JD1 1 dal7D::HIS3 [5] (2) JR85 1 mls1D::LEU2 [5] (3) JR86 2 mls1D::LEU2 dal7D::HIS3 [5] KM10 3 URA3, expressing Mls1pDSKL from the MLS1 locus This study KM11 3 URA3, expressing Mls1p from the MLS1 locus This study (4) BJ1991 MATa leu2 ura3-52 trp1 pep4-3 prb1-1122 gal2 [20] (5) KM12 4 mls1D::LEU2 This study KM13 5 Expressing Mls1pDSKL from the MLS1 locus This study KM15 3 Over-expressing Mls1pDSKL from a multicopy vector This study Plasmids pB10-WT pB10-WTDSKL Plasmid for reintroducing MLS1 at the native locus As above, for introducing an MLS1 truncation This study This study YEp352-MLS1 Multicopy vector harboring native MLS1 [5] YEp352-MLS1DSKL pJR233 Multicopy vector harboring a truncated MLS1 YEp352-based plasmid expressing GFP-SKL This study [22] pJR233M pJR233-derived vector for GFP fusions [21] pLW89 pJR233M-derived plasmid expressing GFP-Mls1p This study Oligonucleotides H161 5¢-CACTGATTTGTGAGAATTCTGATCTCC-3¢ This study H162 5¢-CAATGAACTCTAGAGC-3¢ This study H338 5¢-GATACTAAGTGAGCTTAAGGAGG-3¢ This study H339 5¢-CCCGACGCCGGACGAGCCCGC-3¢ This study H623 5¢-AGAAAGATCTATCTAGTGGGTTGAATTGCGGACGTTGG-3¢ This study H625 5¢-AGAAGCATGCGATCACAATTTGCTCAAATCAGTGGGCGTCGCC-3¢ This study Scheme 2. Diagram of plasmid construction. The pB10-WT or pB10- WTDSKL constructs for expressing M ls1p or Mls1pDSKL f rom th e native locus are shown. Not to scale. PCR oligonucleotide H338 primes 0.25 kb 5¢ of th e PvuII site, H162 primes 0.1 kb 3¢ of the MLS1 ATG start site, H161 primes at a site that includes the MLS1 stop codon, and H339 primes 0 .3 kb 3 ¢ of the Pv uII site. Ó FEBS 2002 Subcellular localization of yeast Mls1p (Eur. J. Biochem. 269) 917 C-terminal 308 a mino acids ( out of a total of 554) was used to express a soluble His-tagged protein (His 6 -Mls1p) in bacterial cells. Cell lysates were subjected to af®nity chromatography using a Ni 2+ -containing Sepharose 6B column (Pharmacia), and protein was puri®ed to near homogeneity using a Ni-nitrilotriacetic acid Spin Kit (Qiagen).SDS/PAGErevealedaproteinbandwithan apparent molecular mass of 38 000, which corresponded t o the d educed size of the His 6 -Mls1p truncation (not shown). A f raction of a puri®ed His 6 -Mls1p was immobilized on a membrane and subjected to tryptic digestion, and HPLC-puri®ed peptide fragments were microsequenced. The sequences obtained, GVHAMGGMAAQIPIK and ATPTDLSK, corresponded to the respective deduced residues 334±348 and 546±553 of Mls1p, con®rming the identity of the puri®ed recombinant protein. The same puri®ed protein (100 lg) in combination with complete Freund's adjuvant (3 mL total volume) was used to immu- nize rabbits (approved by the Ethics C ommittee of the University of Vienna). This was followed by three additional booster injections. After ammonium sulfate precipitation and DEAE-ion exchange of the antiserum, antibody was used for immunoblotting. For immunoelectron microscopy, the antibody preparation w as subjected to af®nity puri®ca- tion using membrane-immobilized soluble protein extracts obtained from yeast cells over-expressing n ative Mls1p. RESULTS The subcellular location of Mls1p Malate synthase 1 terminates with an SKL tripeptide representing a peroxisomal targeting signal P TS1 [5,15]. To determ ine whether Mls1p is i ndeed a peroxisomal protein, electron microscopy was performed using an anti- Mls1p antibody that was g enerated against a recombinant protein comprising the C-terminal 308 amino acids of Mls1p. Although it cannot be entirely ruled out that the antibody used additionally cross-reacts with Dal7p, which is 81% identical to Mls1p and also ends with SKL, expression of Dal7p in cells grown in the presence of ample nitrogen was considered to be unlikely as transcription of the corresponding DAL7 gene is tightly repressed under these medium conditions [5]. Puri®ed antibody was applied t o a ®lter containing soluble protein extracts obtained f rom wild-type a nd mls1D cells that were propagated in rich medium supplemented with ethanol. This resulted i n a protein band with a molecular mass of 62 0 00 in the lane with the wild-type extract that was absent from the lane corresponding to the mls1D mutant (arrow; Fig. 1A), thereby con®rming the speci®city of the antibody. Application of the an tibody to thin se ctions of wild-type cells grown on oleic acid medium Fig. 1. SKL is required to direct Mls1p to the peroxisomes under oleic acid-medium conditions. (A) Speci®city of the anti-Mls1p antibody. Extracts from homogenized wild-typ e (GA1-8C) and mls1D yeast (JR85) strains were immobilized on a membrane to which anti-Mls1p Ig was applied. A single protein band with a molecular mass of 62 000 is seen only in the l ane representing the wild-type extract ( arrow). (B) Immunoelectron mic rograph o f a wild-typ e yeast ce ll expressing native Mls1p from the chromosomal l ocu s (GA1-8C). Gol d particles repr e- senting Mls1p in the matrix of peroxisomes are i ndicated (arrows). l, lipoidal inclu sion; m, mitoch ondrion; n, nucleus; and p, peroxisome. The bar is 1 lm.(C)Micrographofanmls1D mutant over-expressing an SKL-less Mls1p (KM15). Gold particles (marked with arrows) are seen in the nucleus, cytoplasm, and in some case also in mitochondria, peroxisomes, and lipoidal inclusions. The b ar and letters are equivalent to those in ( B). 918 M. Kunze et al.(Eur. J. Biochem. 269) Ó FEBS 2002 resulted in the decoration of peroxisomes ( Fig. 1B). This result lent credence to the suggested peroxisomal location of Mls1p based on a GFP-Mls1p green ¯uorescent protein reporter expressed in cells grown on oleic acid [30]. Use of this antibody with thin sections of an oth erwise i sogenic mls1Ddal7D strain over-expressing an SKL-less Mls1p variant (Mls1pDSKL; strain KM15) o n oleic acid revealed gold particles decorating both the nucleus and cytosol (Fig. 1 C), which was consistent with a non compartmental- ized antigen. The results indicated that the SKL tripeptide was important for peroxisomal targeting. Peroxisomal import of Mls1p depends on oleic acid The glyoxylate cycle is essential for cell growth on media supplemented w ith nonfermentable carbon sources not requiring peroxisomes for their metabolism, e.g. ethanol or acetate, and is physiologically functional in mutant pex cells lacking a normal peroxisomal compartment [19]. This raised the issue o f whether Mls1p is compartmentalized during growth of cells under such medium c onditions. T o examine the subcellular location o f malate synthase 1 in cells grown on ethanol, a GFP reporter was constructed that was extended with the C-terminal 274 amino acids of Mls1p (out of a total of 554), including the terminal SKL. Expression of this GFP-Mls1p was compared to that of a control GFP extended solely by SKL (GFP-S KL). GFP-SKL has been amply shown before to be imported into the peroxisomes of wild-type ce lls, but to remain cytosolic in pex mutant cells devoid of functional peroxisomes [22,31]. The results demonstrated that living yeast cells expressing either GFP- Mls1p o r GFP-SKL on oleic acid exhibited bright, closely bunched ¯uorescent points (Fig. 2, upper panels). On the other hand, in cells grown o n e thanol, t he punctate pattern of ¯uorescence due to GFP-SKL was less dense, whereas ¯uorescence due to GFP-Mls1p was altogether diffuse (Fig. 2 , lo wer panels). This indicated that unlike the situation with GFP-SKL, which was targeted to peroxi- somes in cells grown under both m edium conditions, compartmentalization of GFP-Mls1p into peroxisomes depended on cell growth on oleic acid medium. To reinforce the evidence for the differential subcellular location of Mls1p, cellular fractionation was used. Fractions were prepared from ethanol-grown cells that contained import-competent peroxisomes as they could compartmen- talize GFP-SKL ef®ciently (Fig. 2). Lysates of homogenized wild-type cells were spun to yield an organellar p ellet consisting of mitochondria and peroxisomes, and a cytosolic supernatant. Equal fractions of each of the protein prepa- rations (10% of total vol) were i mmobilized on replicate membranes to which were applied antibodies against Mls1p or yeast peroxisomal Cta1p. The results demonstrated that although Mls1p was c learly detectable in both th e total homogenate and the supernatant (lanes 1 and 2 in the upper panel; Fig. 3A), in the peroxisome-enriched organellar pellet levels of Mls1p w ere below the detection limit (lane 3; Fig. 3A). Cta1p was visible in all three lanes, but was especially abundant in the pellet (lane 3 in t he lower panel; Fig. 3A). Hence, during cell growth under ethanol medium conditions, p eroxisomal Cta1pwas imported, but not Mls1p. Fractionation was also performed on o leic acid-grown cells expressing native Mls1p o r Mls1pDSKL (designated in Fig. 3B as + or ± SKL, respectively). Under these condi- tions, both Mls1p and C ta1p were found in the organellar pellet from cells expressing native Mls1p (lane 5; Fig. 3B). A fairly high proportion of Mls1p and Cta1p was seen in both the s upernatant a nd pellet fractions; it is not yet possible to isolate completely 100% intact organelles. On the other hand, Mls1pDSKL- which could be detected in the homo- genate and s upernatant (lanes 2 and 4) was absent from the corresponding organellar pellet (lane 6). These results con®rmed the requirement of SKL for peroxis omal import, and reiterated that the compartmentalization of malate synthase 1 depended on cell growth on o leic acid medium. Targeting of Mls1p to peroxisomes is advantageous for growth on oleic acid Two steps of the glyoxylate cycle take place in the cytosol: the splitting of isocitrate into succinate a nd glyoxylate, and the dehydrogenation of malate to oxaloacetate (Scheme 1). Fig. 3. Subcellular distribution of native Mls1p under oleic acid- and ethanol medium conditions. (A) Ethanol-grown KM11 cells or (B) oleic acid-grown K M11 and KM10 cells (+ or ±SKL, r espectively) were used for cell fractionation. Aliquots representing 10% of each volume from the primary ho mogenate (hom), the organellar pellet (pellet), o r supernatant (sup) were immobilized to duplicate membranes which were probed with anti-malate synthase (a-Mls1p) or anti-catalase A (a-Cta1p) Ig. Molecular mass markers (kDa) are indicated to the left. Fig. 2. Subcellular localization o f GFP-Mls1p. Oleic a c id-grown BJ1991 c ells transformed with GFP-Mls1p or GFP-SKL were moni- tored by direct ¯uorescence m icroscopy . Punctate ¯uoresc ence indi- cated presence of GFP in peroxisomes. The diuse ¯uorescence seen in ethanol-grown cells expressing GFP-Mls1p was commensurate with a cytosolic localization of t he reporter protein. Nomarski images cor- roborated the integrity of the cells examined. Ó FEBS 2002 Subcellular localization of yeast Mls1p (Eur. J. Biochem. 269) 919 However, the intervening activity undertaken by Mls1p, i.e. formation of malate from g lyoxylate and acetyl-CoA, occurs in the peroxisomes when cells are grown on oleic acid. This prompted the question of whether there is any advantage to cells targeting Mls1p to peroxisomes, as by doing so cells partition the enzyme reactions to either side of the organellar membrane. To examine t he requirement for compartmentalizing Mls1p, yeast mls1D cells (KM12) and strains expressing native Mls1p or Mls1pDSKL from the chromosomal locus (strains KM13 and KM15) were grown on solid fatty acid medium. The medium used also contained Tween 80, which acted to disperse t he fatty acids but was also a poor carbon source. Hence, mutant cells often grow to some extent on these plates but transparent zones in the opaque medium around regions of cell growth indicate utilization of t he fatty acid s ubstrate [25]. Applica- tion of serial dilutions of cell cultures (BJ1991, KM12, KM13) to this medium showed that the mls1D mutant was unable to form a clear zone (Fig. 4A). On the other hand, despite representing a strictly cytosolic protein, Mls1pDSKL appeared to overcome the mutant phenotype (Fig. 4A). To examine whether a cytosolic malate synthase was as ef®cient as a peroxisomal one for m aintaining a functional glyoxylate cycle on oleic acid, liquid growth assays were conducted. The results showed that the growth rate of cells expressing wild-type Mls1p was higher compared with those producing Mls1pDSKL (Fig. 4B). Vital counts based on this assay served to c on®rm that although t he compart- mentalization of malate synthase was not strictly essential, it was advantageous for cells to grow on oleic acid (Fig. 4C). The greater sensitivity of liquid growth assays on oleic acid compared with solid medium has been previously reported [32]. As a control, cells were streaked on eth anol, acetate, or glucose media (Fig. 5A). The results d emonstrated that the mls1D mutant failed to grow on ethanol or acetate. However, expression of either o f the two Mls1p constructs complemented the mls1D mutant phenotype on these media. Growth assays in liquid medium supplemented with ethanol similarly showed that although mls1D cells were unable to multiply, those cells expressing malate synthase in any form, i.e. Mls1p or Mls1pDSKL, grew abundantly (Fig. 5B). This indicated that a constitutively cytosolic Mls1p was suf®cient for cells to maintain the metabolite ¯ux through the glyoxylate cycle during growth o n nonfermentable carbon sources other than fatty acids. DISCUSSION The requirement for the compartmentalization of t he yeast glyoxylate cycle into peroxisomes has been put into question in light of chronicled observations of growth of S. cerevisiae pex mutants devoid of functional peroxisomes on ethanol [19]. I n addition, pex mutants have also been demonstrated to undergo normal meiosis and sporulation in liquid acetate medium [33], p rocesses which similarly require a f unctional glyoxylate cycle [34]. H owever, as pex mutants fail to grow or sporulate in liquid oleic acid medium [33], the issue of the partitioning of the glyoxylate cycle in cells grown under fatty acid-medium conditions has hitherto remained open. We showed here that one of the key glyoxylate-cycle enzymes, Mls1p, was cytosolic in cells grown on ethanol, whereas in cells grown on oleic acid Mls1p was peroxisomal. This is the ®rst t ime that the t argeting of an SKL- terminating protein into peroxisomes is shown to be different depending on the growth conditions. A previous study on the s ubcellular distribution of AKL-terminated Fig. 4. Growth of cells on oleic acid. (A) Plate assay for the utilization of oleic acid. Yeast mls1D cells expressing Mls1p in its native form or without SKL were c ompared with an otherwise i sogenic null mutan t for formation of clear zon es in oleic ac id medium lacking ye ast extract. Strains were grown to late log-phase in rich-glucose medium, and serially diluted culture aliquots were applied to the plates. The plate was r ecorded p hotographically following 5 days incubation at 30 °C. The strains used were BJ1991 (wild type), KM12, and KM13. (B) Cell growth in liquid medium. The strains used were wild type cells (BJ1991, j), mls 1D cells (KM12, r), or mls1D cells complemented with Mls1pDSK L ( KM13, d). The curves represent the average o f three independent experiments. (C) Vital counts of diluted culture aliquots from (B) that were plated on YPD medium. Bars r epresent standard error (n  3). 920 M. Kunze et al.(Eur. J. Biochem. 269) Ó FEBS 2002 aspartate aminotransferase Aat2p demonstrated that this protein was compartmentalized in cells grown on oleic acid, but remained in the cytosol of glucose-grown cells [35]. However, under these latter conditions peroxisomes are very few due to catabolite repression [36,37], whereas on ethanol peroxisomes are not only more readily detectable, but are additionally import co mpetent (Fig. 2). This means that unlike the situation with Aat2p which essentially has no target compartment in cells grown on glucose, Mls1p was selectively retained in the cytosol of cells propagated on ethanol. I nterestingly, t he C-termini of both Mls1p and Aat2p contain acidic amino-acid residues at the 5th-last position with re spect to the terminal residue (DLSKL in Mls1p and EISKL in Aat2p), which i s unusual a t this position [21]. The signi®cance of this similarity is curren tly being addressed. Demonstration of the cytosolic location of Mls1p in wild-type cells grown on ethanol completes the picture of the extra-peroxisomal location of the glyoxylate cycle in yeast grown on carbon sources other t han fatty acids. The only other key enzyme unique to the glyoxylate cycle, Icl1p, is also extra-peroxisomal [4], as are the other enzymes essential for the glyoxylate cycle (Scheme 1) including mitochondrial citrate synthase encoded b y CIT1 (and possibly also by CIT3), cytosolic Mdh2p, and extra- peroxisomal Aco1p. As mentioned previously, malate synthase catalyses the formation of malate from glyoxylate and acetyl-CoA, the source of the latter being either peroxisomal when breaking down fatty acids, or cytosolic when extra-cellular two-carbon substrates are used. Although not strictly essential, the peroxisomal localization of malate synthase 1 appears to be advantageous for cells growing on oleic acid, i n that acetyl- CoA production and u tilization are thereby i ntimately compartmentalized together to increase ef®c iency. Future work on the e ntry of glyoxylate into peroxisomes will help elucidate how the glyoxylate c ycle proceeds a cross an organellar membrane i n cells grown on oleic acid. In addition, solution of the crystal structure of Mls1p could also turn out to be helpful in elucidating whether t he protein's selective import into peroxisomes might have something to do with the e xposure of the C-terminal SKL tripeptide for making contact with t he cognate receptor Pex5p. ACKNOWLEDGEMENTS We dedicate this work to the memory of Professor Helmut Ruis (University of Vienna), who p assed away unexpectedly on September 1st 2001, aged 61. We t hank Jana Raupadioux and, Leila Wabnegger for e xcellent technical assistance . Dr Hanspeter Rottensteiner (FU Berlin, Germany) and Professor J. Kalervo Hiltunen (University of Oulu, Finland) are gratefully acknowledged for useful suggestions. The work was supported by the Fonds zur Fo È rderung der wissenschaftli- chen Forschung (FWF), Vienna, Austria (grants P9398-MOB and P12118-MOB to A. H.). REFERENCES 1. de Duve, C. & Baudhuin, P. (1966) Peroxisomes (microbodies and related particles). Physio l. Rev. 46, 323±357. 2. Kunau, W H., Dommes, V . & Schulz, H. (1995) b-Oxidation of fatty acids i n mitochondria, peroxisomes, and bacteria: a century of continued progress. Prog. Lipid Res. 34, 267±342. 3. Lorenz, M.C. & Fink, G.R. (2001) The glyoxylate cycle is required for fungal virulence. Nature 412, 8 3±86. 4. Taylor, K.M., K aplan, C.P., Gao, X. & Baker, A. (1996) Local- ization and targeting of isocitrate l yases in Saccharomyces cerevi- siae. Biochem. J. 319, 255±262. Fig. 5. Growth of cells on e thanol. (A) Plate assays for functiona l complementation of a yeast mls1D strain (JR86) expressing native Mls1p (KM11) or an SKL-less variant (KM10) on ethanol, acetate, or glucose media, as indicated. (B) Cell growth in liquid ethanol medium. The strains used were identical to those in Fig. 4. The curves represent the average of three independent experiments. Ó FEBS 2002 Subcellular localization of yeast Mls1p (Eur. J. Biochem. 269) 921 5. Hartig,A.,Simon,M.M.,Schuster,T.,Daugherty,J.R.,Yoo, H.S. & Cooper, T.G. (1992) Di erentially regulated malate s yn- thase genes participate in carbon and nitrogen metabolism of S. cerevisiae. Nucleic Acids Res. 20, 5677±5686. 6. McAlister-Henn, L. & Thompson, L.M. (1987) Isolation and expression of th e g ene e ncoding y e ast mito chondrial m alate dehydrogenase. J. Bacteriol. 169, 5157±5166. 7. Minard, K.I. & McAlister-Henn, L. (1991) Isolation, nuc leotide sequence analysis, and disruptio n of the MDH2 gene from Sac- charomyc es cerevisiae: evidence for three isozymes of yeast malate dehydrogenase. Mol. Cell. Biol. 11, 370±380. 8. Stean, J .S. & McAlister-Henn, L. (1992) Isolation and charac- terization of the yeast gene enco ding th e MDH3 isozyme of malate dehydrogenase. J. Biol. C hem. 267, 2 4708±24715. 9. Van Roermund, C.W., Elgersma, Y., Singh, N., Wanders, R.J. & Tabak, H.F. (1995) The membrane of peroxisomes in Sa cchar- omyces cerevisiae is impermeable to N AD (H) a nd acetyl- CoA under in vivo conditions. EMBO J. 14, 3480±3486. 10. Suissa, M., Suda, K. & S chatz, G. (1984) Isolation of the nuclear yeast g enes for citrate syn thase and ®fteen oth er mitochondrial proteins by a n ew screening method. EMBO J. 3, 1773±1781. 11. Kim, K.S., Rosenkrantz, M.S. & Guarente, L . (1986) Sacchar- omyces cerevisiae contains two functional citrate synth ase genes. Mol. Cell. Biol. 6, 1936±1942. 12. Lewin, A.S., Hines, V. & Small, G.M. (1990) Citrate synthase encoded by t he CIT2 gene of Saccharomyces cerevisiae is perox i- somal. Mol. Cell. Biol. 10, 1399±1405. 13. Przybyla-Zawislak, B., G adde, D.M., Ducharme, K. & McCammon, M.T. ( 1999) G enetic and biochemical interactions involving tricarb oxylic acid cycle ( TCA) func tion using a c ollec- tion of mutants defective in all TCA cycle genes. Genetics 152, 153±166. 14. Ganglo, S.P., M arguet, D . & Lauquin, G.J . (1990) M olecular cloning of the yeast mitochond rial aconitase ge ne (ACO1)and evidence of a synergistic regulation of expression by glucose p lus glutamate. Mo l. Ce ll. Bio l . 10, 3551±3561. 15. Gould, S.J., Keller, G A. & Subramani, S. (198 7) Identi®cation o f a peroxisomal targeting signal at the carboxy terminus of ®re¯y lucifera se . J. Cell Biol. 105, 2923±2931. 16. Gould, S.J., Keller, G A., Hosken, N., Wilkinson, J. & Subra- mani, S. (1989) A c onserved tripeptide sorts protei ns to pe roxi- somes. J. Cell Biol. 108, 1657±1664. 17. Gould, S.J., Keller, G A., Schneider, M., Howell, S.H., Garrard, L.J., Goodman, J.M., Diste l, B., T abak, H.F. & S ubramani, S. (1990) Peroxisomal protein import is conserved between yeast, plants, insects and mammals. EMBO J. 9, 85±90. 18. Elgersma, Y., Van Roermund, C.W., Wanders, R.J. & Tabak, H.F. (1995) Peroxisomal and mitochondrial carnitine acetyl- transferases of Saccharomyces cerevisiae are encoded by a single gene. EMBO J. 14, 3472±3479. 19. Erdmann, R., Veenhuis, M., Mertens, D. & Kunau, W H. (1989) Isolation of peroxisome-de®cient mutants o f Saccharomyces cerevisiae. Proc. Natl A cad. Sci. USA 86, 5419±5423. 20. Jones, E.W. (1977) Proteinase mutants of Saccharomyces cerevi- siae. Genetics 85, 23±33. 21. Lametschwandtner, G., Brocard, C., Fransen, M., Van Veldho- ven, P., B erger, J. & Hartig, A. (1998) The dierence in recognition of terminal tripeptides as peroxisomal targeting signal 1 between yeast and human is due to dierent anities of their receptor Pex5p to the cognate signal and to residues adjacent to it. J. Biol. Chem. 273, 33635±33643. 22. Brocard, C., Lametschwandtner, G., Koudelka, R. & Hartig, A. (1997) Pex14p is a member of the protein linkage map of Pex5p. EMBO J. 16 , 5491±5500. 23. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: a Laboratory M anual, 2nd edn. Cold Spring Harbor Laboratory Pres s, Cold Spring Harbor , NY. 24. Chen, D C., Yang, B C. & Kuo, T T. (1992) One-step t rans- formation of yeast in stationary phase. Curr. Genet. 21, 83±84. 25. Gurvitz, A., Rottensteiner, H., Kilpela È inen, S.H., Hartig, A., Hiltunen, J.K., Binder, M., Dawes, I.W. & Hamilton, B. (1997) The Saccharomyces cerevisiae peroxisomal 2,4-dienoyl-CoA reductase is enc oded by t he oleate -inducib le gene Sps19 . J. Biol. Chem. 272, 22140±22147. 26. Gurvitz, A., Mursula, A.M., Firzinger, A., Hamilton, B., Kil- pela È inen, S.H., Hartig, A., Ruis, H., Hiltunen, J.K. & Rottenste- iner, H. (1998) Peroxisomal D 3 -cis-D 2 -trans-enoyl-CoA isomerase encoded by ECI1 is required for growth of the yeast Sacchar- omyces cerevisiae on unsaturated f atty acids. J. Biol. Chem. 273, 31366±31374. 27. Kragler, F., Langeder, A., Raupachova, J., Binder, M. & H artig, A. (1993) Two independent peroxisomal targeting signals in catalase A of Saccharomyces cerevisiae. J. Cell Biol. 120 , 665±673. 28. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the h ead of bacteriophage T4. Nature 227, 680±685. 29. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of micro gram quantities o f protein u tilizi ng the principle of protein-dye binding. Anal. Biochem. 72, 248±254. 30. Geraghty,M.T.,Bassett,D.,Morrell,J.C.,Gatto,G.J.Jr,,Bai,J., Geisbrecht, B.V., Hie ter, P . & Go uld, S.J . (19 99) De tecting p at- terns of protein distribution and gene expression in silico. Proc . Natl Acad. Sci. USA 96 , 2937±2942. 31. Monosov, E.Z., Wenzel, T.J., Lu È ers, G.H., Heyman, J.A . & Subramani, S. (1996) Labeling of pe roxisome s with green ¯uo- rescent protein in living P. pastoris cells. J. Histochem. Cytochem. 44, 581±589. 32. Qin, Y M., Marttila, M.S., Haapalainen, A.M., Siivari, K.M., Glumo, T. & Hiltunen, J.K. (1999) Yeast peroxisomal multifunctional enzyme: (3R)-h ydroxyac yl-CoA dehy drogenase domains A and B are required for optimal gro wth on oleic acid. J. Biol. Chem. 274, 28619±28625. 33. Gurvitz, A., Rottensteiner, H., Hamilton, B., Ruis, H., Hartig, A., Dawes, I.W. & Binder, M. (1998) Fate and r ole of peroxisomes during t he life cycle of the yeast Saccharomyces cerevisiae: inher- itance of peroxisomes during m eiosis. Histochem. Cell Biol. 110, 15±26. 34. Dickinson, J.R., Dawes, I.W., Boyd, A.S. & Baxte r, R .L. (1983) 13 C NMR studies of acetate metabolism during sporulation of Saccharomyces cerevisiae. Proc.NatlAcad.Sci.USA80, 5847± 5851. 35. Verleur, N., Elgersma, Y., Van Roermund, C.W., Tabak, H.F. & Wanders, R .J. (1997) Cytosolic aspartate aminotransferase encoded by the AAT2 gene is targeted to the peroxisomes in oleate-grown Saccharomyces cerevisiae. Eur. J. Biochem. 247, 972±980. 36. Veenhuis, M., Mateblowski, M., Kunau, W H. & Harder, W. (1987) Proliferation of microbodies in Saccharomyces cerevisiae. Yeast 3, 77±84. 37. Rottensteiner, H., Kal, A.J., Filipits, M., Binder, M ., Hamilton, B.,Tabak,H.F.&Ruis,H.(1996)Pip2p:atranscriptionalregu- lator of p eroxisome proliferation in the yeast Saccharomyces cerevisiae. EMBO J. 15, 2924±2934. 922 M. Kunze et al.(Eur. J. Biochem. 269) Ó FEBS 2002 . Targeting of malate synthase 1 to the peroxisomes of Saccharomyces cerevisiae cells depends on growth on oleic acid medium Markus Kunze 1 , Friedrich. additionally representsthesolesiteforfattyacidb-oxidation. The sub- cellular location of the key glyoxylate-c ycle enzyme malate synthase 1 (Mls1p), an